←  All terms
C
Performance

Cumulative Layout Shift

Cumulative Layout Shift measures unexpected movement of visible elements while a web page is being used.

Cumulative Layout Shift measures unexpected movement of visible elements while a web page is being used. Optimization becomes meaningful when real-user data, repeatable lab tests, and an explicit performance budget are used together. The concept is closely related to Core Web Vitals and Image Optimization.

Read more

Core Web Vitals

Image Optimization

CSS

FAQ

Cumulative Layout Shift

Does image optimization alone prevent layout shifts caused by images?
No, because compression shrinks files but does not reserve space. A small image without dimensions still pushes the text below it once it loads. Both are needed: lean formats for load time and fixed dimensions or an aspect ratio for a stable layout. This also applies to responsive image variants.
Why do lab and field data differ for CLS?
Lab tools usually measure only the initial load of a simulated visit. Field data also captures scrolling, late-loading content, consent banners, and varied devices. A lab test can therefore show 0 while real users experience shifts. Core Web Vitals assessment relies on field data, while the lab serves for debugging.
How can layout shifts be avoided in practice?
Reserve space before content arrives. That means width and height or a fixed aspect ratio for media, fixed containers for embeds and banners, and matching fallback fonts. New content belongs below the visible area. Animations should use transforms rather than properties that force the layout to be recalculated.
What typically causes layout shifts on web pages?
Usually reserved space is missing: images without dimensions, late-loading ad and embed slots, cookie banners that push content down, and web fonts with different letter metrics. Shifts right after a user action are not counted. What matters is whether content jumps unexpectedly, not how much loads late.
At what value is Cumulative Layout Shift considered good?
A CLS value of 0.1 or lower counts as good. The benchmark is the 75th percentile of field data, meaning the value that three out of four page views stay below. Single lab runs give only a hint. When the threshold is missed, the largest individual shifts should be identified first rather than improving an average.